Transcript

ANALYSIS OF TAPERED LAMINATED COMPOSITE TUBES UNDER TENSION

AND TORSION

by

CHETHANA SHANKARA RAO

Presented to the Faculty of the Graduate School of

The University of Texas at Arlington in Partial Fulfillment

of the Requirements

for the Degree of

MASTER OF SCIENCE IN MECHANICAL ENGINEERING

THE UNIVERSITY OF TEXAS AT ARLINGTON

May 2007

ACKNOWLEDGMENTS

I would like to thank Dr. Chan for his patience and time, and his unlimited

guidance and support, which made this thesis possible. I would also like to thank Dr.

Lawrence and Dr. Nomura for serving on my thesis committee. I would also like to thank

all the professors of the MAE department because of whom; my master’s degree has been

a huge success in terms of learning and knowledge acquisition.

I would also like to express my gratitude towards my parents, my beloved brother

and my friends for their unending support and encouragement throughout my career,

especially during my master’s degree.

April 20th , 2007

ii

ABSTRACT

ANALYSIS OF TAPERED LAMINATED COMPOSITE TUBES UNDER TENSION

AND TORSION

Publication No. ______

Chethana Shankara Rao, M.S.

The University of Texas at Arlington, 2007

Supervising Professor: Wen S Chan

Closed form expressions for determining the displacement and twisting angle of

tapered composite tubes are developed. The analytical expressions are developed based

on the modified laminated plate theory which includes the tubular wall curvature of the

laminate. It is found that the axial deformation and the twisting angle calculated by the

current method agree well with the results obtained from the finite element method.

The effect of stacking sequence, taper angle and fiber orientations on the axial

deformation and twisting angle are studied by using the developed method. It is found

that the taper angle plays a significant role on the axial deformation and the twisting

angle provided the laminate lay-up is given.

iii

TABLE OF CONTENTS

ACKNOWLEDGMENTS...................................................................................................ii

ABSTRACT.......................................................................................................................iii

LIST OF ILLUSTRATIONS...........................................................................................viii

LIST OF TABLES..............................................................................................................x Chapter 1. INTRODUCTION .............................................................................................. 1

1.1 Overview ………………………………………………………………………1

1.2 Literature Survey ……………….……………………………………………...4

1.3 Objective and Approach of Thesis…..………………………………………….7

1.4 Outline of the Thesis ………………………………………………..………….8

2. STIFFNESS MATRICES OF UNIFORM CIRCULAR TUBES….……...………9

2.1 Geometry of the Tube ……….………………….……………………………..9 2.2 Stiffness Matrices of the Tube ……..……………………….…………………9

2.3 Axial Stiffness………………………………………...………………………14

2.4 Torsional Stiffness…………………………………………………………….14

2.5 Smeared Property Approach…………………………………………………..15

2.6 Bending Stiffness Comparison between the Approaches …………………….15

3. TAPERED TUBES UNDER AXIAL TENSION………………...…..……...…..16

3.1 Analytic Solution for an Isotropic Tube….…………………………………...16

3.2 Finite Element Model of an Isotropic Tube....……….…………….…....…….18

iv

3.2.1 Geometry and Modeling………………………..……………………....18

3.2.2 Development of the Model….……………………………………….....19

3.2.3 Boundary Conditions……….…………………………………………..20

3.3 Results Comparison….…………………………………………….…………22

3.4 Analytical Model of the Composite Tube …………………………..……..…23

3.4.1 Geometry………………………...……………………………………..23

3.4.2 Stiffness Matrices………..……………………………………………..23

3.4.3 Axial stiffness and Stress Distribution….…………..………………….26

3.4.4 Calculation of the Stresses from Analytical Solution…………………..28

3.5 Finite Element Model of the Composite Tapered Tube……….……………...29

3.5.1 Geometry of the tube…………………………………………………….29

3.5.2 Meshing the model……………………………………………………….29

3.5.2.1 Element Description………………………………………………29

3.5.2.2 Generation of the Model…………………..………………………30

3.5.3 Boundary Conditions………….………………………………………….31

3.6 Results Comparison……………...……………………………………………31

3.6.1 Comparison of Axial Stresses…..….…………………………………….35

4. TAPERED TUBES UNDER TORSION……...…………….…………………... 36

4.1 Analytical Solution for an Isotropic Tube….…………………………………36

4.2 Finite Element Model of an Isotropic Tube…...……….…...….……....……...37

4.2.1 Meshing the Model……....………………………………………………38

4.3 Results Comparison…………….………..……………………………………39

v

4.4 Analytical Model of the Composite Tube…….………..……………………..40

4.4.1 Stress Calculations……………..……………………………………..42

4.5 Finite Element Method of the Composite Tube ……..………………………..42

4.5.1 Post Processing and Results Comparison…………………………….42

4.5.2 Comparison of Shear Stresses ( xyτ )………………………………….47

5. PARAMETRIC STUDIES.................…………………………………..………….48

5.1 Stacking Sequence Effect………..…………………………………………...48

5.1.1 Axial Case………………………………………………………….....48

5.1.2 Torsion Case………….……………………………………………….49

5.2 Effect of Fiber Orientation ………………………………………...…………50

5.2.1 Axial Case……………………………………………………………..50

5.2.2 Torsion Case………………………………………….………………..52

5.3 Effect of the Taper Angle…………..…………………………...…………….54

5.3.1 Axial Case……………………………………………………………..54

5.3.2 Torsion Case…………………...………………………………………55

6 CONCLUSIONS………………………………………………………………...........57 Appendix

A. TERMS OF THE STIFFNESS MATRICES……………………….…………….58

B. MATHEMATICA CODE FOR THE ANALYTICAL SOLUTION OF AXIAL DEFORMATION OF AN ISOTROPIC TUBE……...…………………61

C. MATHEMATICA CODE FOR THE ANALYTICAL SOLUTION OF

THE ANGLE OF TWIST OF AN ISOTROPIC TUBE……...………….............63 D. MATHEMATICA CODE FOR THE ANALYTICAL SOLUTION OF

AXIAL DEFORMATION AND ANGLE OF TWIST OF A

vi

TAPERED COMPOSITE TUBE……………..………..………………………..65 E. ANSYS BATCH CODE FILE FOR THE GENERATION OF THE

FINITE ELEMENT MODEL…….…………………………..……….………….68 REFERENCES………...…….………………………………………………………..…73

BIOGRAPHICAL INFORMATION…..…………………..………...……………..……77

vii

LIST OF ILLUSTRATIONS

Figure 1.1 Composite baseball bat…………………………………………………………2 1.2 Composite tennis rackets……………………………………………………….3 1.3 Composite golf shaft………………………………………………….……….. 3 1.4 Composite CNG tank……………………………………….…………….…….4

2.1 Development of Analytic Solution using Laminated Plate Theory.....…….….13

3.1 Dimensions of the Tapered Tube...……………........….…………...………....16

3.2 Development of the Finite Element Model ...….......…………………....…….19

3.3 Convergence Plot…………………………………….......................................21

3.4 Finite Element Model of an Isotopic Tube with Boundary Conditions…...…..21

3.5 Graph of Normalized Displacement Vs Distance from Fixed End......………..23

3.6 Geometry of the Tapered Composite Tube …...…...….………………..……..24

3.7 Development of the Analytical Model from the Laminated Plate Theory .…...25

3.8 Flow Chart of the Development of the Analytical model…………….….…... 26

3.9 Solid 191 Element Geometry ……......…...…..………………...……………..30

3.10 Graph of axial deformation vs. distance from fixed end

for the composite tube…..…………………………………………………….32

3.11 Plot showing the displacement of the tip of the tapered tube…….……………33

3.12 Stresses in the direction in the ply.…..…………....…………...………33 x 045+

3.13 Stresses in the direction in the ply ….…….…..…………………..….. 34 x 045−

viii

3.14 Stresses in the direction in the ply .........…..……….…………………… 34 x 00

4.1 Finite Element Model of a Tapered Isotropic Tube under Torsion with Boundary Conditions....…...…………………………………………….……..38

4.2 Graph showing normalized angle of twist from analytical and finite

element method Vs distance from fixed end….………………………..………40

4.3 Calculation of the Angle of twist from Finite Element Model.….…...……...…43

4.4 Graph of normalized angle of twist vs. Distance from fixed End.…......………45

4.5 Shear stresses on the ply… ……...…………………………………...…..45 045+

4.6 Shear stresses on the ply………………………...……………………….46 045−

4.7 Shear stresses on the ply……………………......…………………………..46 00

5.1 Variation of axial deformation with fiber orientation...…..……………..……..51 .

5.2 Variation of angle of twist with fiber orientation ...…..……..……………..…..53

5.3 Equivalent Shear modulus plot...………………………………………………..53

5.4 Variation of normalized axial deformation with taper angle................................55

5.5 Variation of normalized angle of twist with taper angle….......……………. …..56

ix

LIST OF TABLES

Table

3.1 Table showing the convergence of results with increase in the number of elements………………..……..………………………………………….…….20

3.2 Comparison of normalized axial deformation from analytical method and

finite element method…...…………………………………………………….22

3.3 Comparison of normalized axial deformation from analytical method and finite element method for a composite tube.………..….……………………..32

3.4 Comparison of stresses................………….…………………………………..35

4.1 Comparison of angle of twist results from analytical method and finite

element method for an isotropic tube………..………………………………..39

4.2 Comparison of angle of twist results from analytical method and finite element method for a composite tube………..………………………………..44

4.3 Comparison of Shear stresses…....……………………………………………47

5.1 Variation of axial deformation with stacking sequence……..………………..49

5.2 Variation of angle of twist with stacking sequence …...…...…...….…………49

5.3 Variation of axial deformation with fiber orientation..….......…...………..…..51

5.4 Variation of angle of twist with fiber orientation….......………..…………… 52

5.5 Variation of axial deformation with taper angle α ..………...…………....…..54

5.6 Variation of angle of twist with taper angle α …………………………..……56

x

CHAPTER 1

INTRODUCTION

1.1 Overview

The applications of composite materials in aeronautical and other engineering

structures are ever increasing, due to their highly desirable properties like high specific

strength/ stiffness, low co-efficient of expansion, damping properties and directional

dependence. One of the first uses of composite materials was about 30 years ago in the

empennage skins of F14 aircraft. With extensive research in the field, composite

materials are now being applied to primary structures of many aircrafts including

fuselage and wings. Of-late composites are gaining popularity in automobile and civil

industries. Fiber reinforced plastics (FRP), which is one of the most widely used

composite in automobile industry, contributes towards reduced fuel consumption,

increased pay load, strength and stiffness and lesser corrosion.

Laminated composite shells in particular are prominent in bearing various types

of loads and are hence used in many engineering structures. The analysis of laminated

composites is quite complicated since the material behavior is anisotropic, which gets

further intensified in the analysis of complex structures like shells. Laminated composite

shells are used in fuselage structures, pressure vessels, missiles and spacecraft, jet nozzles

etc. A shell is defined as a thin walled body with a curvature at least in one direction.

1

The use of thin walled tube constitutes the most unified approach for

characterizing the mechanical properties of fiber reinforced composites. A thin walled

tube, whose radius to thickness ratio is very large, is under a state of uniform stress under

combined tension, torsion and internal pressure. Some applications of composite shells

are described below.

Carbon fiber baseball bats (see Fig 1.1) are used in amateur baseball since the

flying distance becomes longer and serves as a good replacement for wood, which is is

becoming less and less available.

Fig 1.1 Composite baseball bat

Tennis rackets (Fig 1.2) made of carbon fibers are far lighter than wooden rackets

and have higher service life than wooden rackets.

2

Fig 1.2 Composite tennis rackets

Golf shafts (Fig 1.3) made of composites are much lighter than metal shafts

giving higher swing speed for a given power. Composite golf shafts, which were used

initially only by senior players and ladies is now gaining popularity among professional

and amateur players.

Fig 1.3 Composite golf shaft

3

Composites are used in compressed natural gas (CNG) tanks (Fig 1.4) for

natural gas cars due to their light weight.

Fig 1.4 Composite CNG tank

Other applications include wind mill blades, rollers, shafts and X ray inspection

equipment. Laminated cylindrical shells are used in pressure vessels, missiles and

aircraft, fuselage structures etc. Laminated conical shells are often used as transition

elements between cylinders of different diameters, end closures etc. The use of laminated

composites in conical shells are however limited to tubes of low radius ratio, since high

radius ratios results in a non-uniform orientation of fibers in a ply.

1.2 Literature Survey

Good research has been carried out in the past; many of which focus on the

behavior of laminated shells under different loading conditions, hygrothermal effects etc.

Other studies in the area include buckling and post buckling analysis of laminated shells,

delaminations and thermal analysis. Many of these studies have been carried out using

finite element technique.

4

Reddy [1] gives a detailed description of the analysis of laminated plates in

vibration; buckling and bending using the Classical Laminated Plate theory and First

order shear deformation theory. He also discusses the vibration and buckling

phenomenon of cross ply laminated circular cylindrical shells.

Much focus has not been poured on the analytical modeling of laminated shells.

While finite element techniques are cumbersome and time consuming, experimental

techniques need proper specimens and mounting techniques. In such conditions, there is a

high need for an easy-to-use analytical solution. Demirhan and Chan [2] have presented

two analytical closed form expressions for the evaluation of the stiffness matrices of

cylindrical composite tube, one employing the laminated plate theory and the other

employing the laminated shell theory. The models are compared with the smear property

approach and also with the results from a finite element model. In their model, effect of

curvature of the laminate in tubes are included Recently, Lin and Chan [3] modified the

former model for circular tube to develop a model for elliptical cylindrical tube under

bending. Finite element model is used for the verification of this model.

Ren [4] has developed and elastic solution for an anisotropic laminated circular

cylindrical shell simply supported and subjected to axisymmetric loads using three

dimensional elasticity theory. Many research has been conducted on laminated shells

using numerical techniques like finite element method. Tafreshi [5] has presented a

computational model for the delamination in isotropic and laminated composite

cylindrical shells. A combined double layer and single layer of shell elements are

5

employed to study the effects of delamination on the global load carrying capacity of

such systems under axial compressive load. Vaziri [6] has studied the sensitivity of

buckling behavior of cracked composite cylindrical shells to service life buckling by

performing a linear buckling analysis. Computational models are developed by using a

special meshing technique in which the element size is reduced incrementally from the

uncracked region to the cracked tip. Gummadi and Palazotto [8] have presented a

progressive failure analysis on cylindrical shells. Non linear finite element formulation

with large rotation capability is used to predict the onset of failure modes. The different

failure modes considered are: fiber breakage, matrix cracking and delamination.

Yan, Ying and Chen [7] have studied the behavior of simply supported laminated

cylindrical shell with viscoelastic interfaces in cylindrical bending.

Very little work has been carried out in the field of laminated conical shells. Most

of them deal with buckling and vibration, some of which are presented below.

Correia, Soares and Herskovits [9] have presented a numerical method for the

structural analysis of laminated conical shells using a quadrilateral isoparametric element

based on higher order shear deformation theory. The model developed can be used to

perform static analysis with arbitrary boundary conditions and loads and for solving eigen

values problems.

Liu [10] has developed a theory for non linear bending of symmetrically

laminated, cylindrically orthotropic, shallow conical shells subjected to an

6

axisymmetrically distributed load. Transverse shear effects were also included in his

model.

Goldfield [11] has studied the sensitivity of stiffened conical shells to

imperfection via post buckling analysis.

Raju, Chandra and Rao [12] have presented the transient temperatures in

laminated composite conical shells subjected to aerodynamic heating. Unsteady heat

conduction equations for a laminated composite conical shell corresponding to an

axisymmetric temperature field are formulated.

Most of the research in laminated conical shells is based on thermal and buckling

analysis. No attempt has been made in the past to study the structural response of

laminated conical shells (tapered composite tubes).

1.3 Objective and Approach of Thesis

In the current research, an effort is made to develop an analytical model for the

computation of deformations of tapered laminated composite tube under axial tension and

torsion. IM6/ 3501-6 Graphite/ Epoxy is used as the composite material in the current

research. The analytical models are developed based on the laminated plate theory and

are extensions of the model developed for circular cylindrical tubes [3], which is

discussed in detail in Chapter 2. The analytical model developed is more like a procedure

than a closed form solution. The models are validated using a finite element model,

7

which is first used to analyze an isotropic tube in order to ensure the validity of the

model. The finite element model is developed in the computer program ANSYS.

Parametric studies are conducted on tubes to study the effect of stacking sequence, fiber

orientation and taper angles.

1.4 Outline of the Thesis

A review on circular cylindrical tubes [2] is carried out in Chapter 2. Analytical

solution and finite element model for tapered tubes (Isotropic and composite) under axial

tension are discussed in Chapter 3. Chapter 4 deals with analytical model and finite

element model of tapered tubes (Isotropic and Composite) under torsion. Parametric

studies to study the effects of stacking sequence, fiber orientation and taper angles are

discussed in Chapter 5. Finally conclusions and recommendations are presented in

Chapter 6. MATHEMATICA codes and ANSYS batch files are presented in the

Appendix section.

8

CHAPTER 2

STIFFNESS MATRICES OF UNIFORM CIRCULAR TUBES

Chan and Demirhan [2] developed two new analytical methods for evaluating the

stiffness matrices of laminated uniform circular tubes. The two methods laminated plate

approach and laminated shell approach account for the stacking sequence and curvatures.

The laminated shell approach is however not relevant, hence not discussed here. Both the

theories are developed based on the lamination theory.

2.1 Geometry of the Tube

The laminated tube considered is a uniform tube, with circular cross section with

an outer radius oR , inner radius iR and a length L . The length of the tube is sufficiently

larger than its radii. Hence the tube considered is a long tube. In all the derivations the

basic assumption of ‘Plane remains plane after deformation’ is made.

2.2 Stiffness Matrices of the Tube

The current model called the laminated plate theory is based on the conventional

lamination theory. According to the lamination theory the stiffness matrices of an

laminate with a ply thickness are given by:

n ply

plyt

9

[ ] ( )

[ ] (

[ ]

)

( )

11

2 21

1

3 31

1

12

13

n

k kk k

n

k kk k

n

k kk k

A Q h h

B Q h h

D Q h h

−=

−=

−=

⎡ ⎤= −⎣ ⎦

⎡ ⎤= ⎣ ⎦

⎡ ⎤= −⎣ ⎦

∑

∑

∑

− (2.1)

where k

Q⎡ ⎤⎣ ⎦ is the reduced stiffness matrix of the layer. In the above form, the thkk

Q⎡ ⎤⎣ ⎦

matrices have transformation matrices in the direction only, which accounts for the

fiber orientation. In the current research, along with a transformation about , we also

have to include a transformation about the direction, which is discussed later in the

chapter.

z

z

x

In order to develop the model, an infinitesimal plate element of the tube is

considered as shown in the Fig. 2.1. The infinitesimal element is inclined at an angle θ to

the axis of the tube. In order to develop the model, reduced stiffness matrices of the

plies, which accounts for the fiber orientation are evaluated.

z′

( ) [ ] ( )'z z

Q T Q Tσ εα α⎡ ⎤ ⎡ ⎤ ⎡= − ⋅ ⋅ + ⎤⎣ ⎦ ⎣⎣ ⎦ ⎦

The element which is inclined at an angle θ with the z′ axis is rotated about

axis to make it parallel to x′ y′axis. Hence the reduced stiffness matrix, [ ]Q′ of the

individual plies needs to be transformed about the axis. As we know, the transformation x

10

of a reduced stiffness matrix, has both stress and strain transformation

matrices,[ ]Tσ

and[ ]Tε

. These matrices are obtained as described below.

Since, the infinitesimal element is dimensional, we can get the stress

transformation matrix [

2

]Tσ

for a case by striking off the 2, 3 and 42D th rows and

columns of the 3 dimensional [ ]xTσ matrix.

Thus, [ ] 2

1 0 0

0

0 0x

T m

mσ 0

⎡ ⎤⎢ ⎥

= ⎢ ⎥⎢ ⎥⎢ ⎥⎣ ⎦

(2.2)

Similarly we can find that the strain transformation matrix, [ ]Tε

is equal to the stress

transformation matrix [ ]Tσ

. Hence,

[ ] 2

1 0 0

0 0

0 0x

T m

mε

⎡ ⎤⎢ ⎥

= ⎢ ⎥⎢ ⎥⎢ ⎥⎣ ⎦

We can also note that, since the transformation matrices contain only , the cosine

terms, the matrices are the same for

m

θ+ or θ− terms. Hence, reduced stiffness matrix

after rotation will be given by: x

[ ] [ ]'x

Q T Q Tσ⎡ ⎤ ⎡ ⎤= ⎣ ⎦⎣ ⎦ xε (2.3)

The stiffness matrices,[ ]A , [ and ]B [ ]D are then calculated using the lamination theory:

11

[ ]

[ ]

[ ]

11

2 21

1

3 31

1

( )

1 (2

1 ( )3

n

k kk k

n

k kk k

n

k kk k

A Q h h

)B Q h h

D Q h h

−=

−=

−=

⎡ ⎤= −⎣ ⎦

⎡ ⎤= ⎣ ⎦

⎡ ⎤= −⎣ ⎦

∑

∑

∑

− (2.4)

The above matrices need to be translated to the y′axis using the parallel axis theorem.

Here, z R Cosθ= . Hence, the new stiffness matrices are given by:

[ ]

[ ] [ ]

[ ] [ ] ( ) [ ]

'

'

2' 2

A A

B B R Cos A

D D R Cos B R Cos

θ

θ θ

⎡ ⎤ =⎣ ⎦

⎡ ⎤ = +⎣ ⎦

⎡ ⎤ = + +⎣ ⎦ A

(2.5)

Note that the stiffness matrices are in terms of R , where R is the mid-thickness radius of

the circular tube, 0 2 plynR R t= − .

These matrices are then integrated over the entire θ domain to evaluate the stiffness

matrices of the tube at the section.

2'

0

2'

0

2'

0

A R A d

B R B d

D R D d

π

π

π

θ

θ

θ

⎡ ⎤⎡ ⎤ =⎣ ⎦ ⎣ ⎦

⎡ ⎤⎡ ⎤ =⎣ ⎦ ⎣ ⎦

⎡ ⎤⎡ ⎤ =⎣ ⎦ ⎣ ⎦

∫

∫

∫

(2.6)

Note that R is not a function of θ and is hence taken out of the integration.

12

Fig 2.1 Development of Analytic Solution using Laminated Plate Theory

The above matrices can be expanded as:

[ ]

[ ] [ ]

[ ] [ ] [ ]

2

0

2 22

0 0

2 2 22 3

0 0 0

2

A R A d

B R B d R A Cos d

D R D d R B Cos d R A Cos d

π

π π

π π π

θ

θ θ θ

2θ θ θ θ θ

⎡ ⎤ =⎣ ⎦

⎡ ⎤ = +⎣ ⎦

⎡ ⎤ = + +⎣ ⎦

∫

∫ ∫

∫ ∫ ∫

(2.7)

13

2.3 Axial Stiffness

The relationship between forces and moments on the tube and the mid-plane

strains and curvatures are given by:

oN A B

M B D

ε

κ

⎡ ⎤ ⎡ ⎤⎡ ⎤⎢ ⎥= ⋅ ⎢ ⎥⎢ ⎥⎢ ⎥⎢ ⎥ ⎢ ⎥⎣ ⎦ ⎣ ⎦⎣ ⎦

o

T

a b N

Mb d

ε

κ

⎡ ⎤⎡ ⎤ ⎡ ⎤⎢ ⎥= ⋅⎢ ⎥ ⎢ ⎥⎢ ⎥ ⎢ ⎥⎢ ⎥ ⎣ ⎦⎣ ⎦ ⎣ ⎦

(2.8)

oε and are the mid-plane strains and curvatures. κ

For an axial tensile case, is the only applied load. Hence, the mid-plane axial

deformation

xN

xoε is given by:

11xo a Nε x= ⋅ (2.9)

Hence, the axial stiffness will be given by 11

1a

2.4 Torsional Stiffness

For a torsion case, only a moment xyM (torque T ) is applied. Hence the mid-

plane angle of twist will be given by, xyκ

66xy xyd Mκ = ⋅ (2.10)

Hence the torsional stiffness is given by 66

1d

14

2.5 Smeared Property Approach

This approach is more often used for the calculation of extensional and bending

stiffness of the tube. The equivalent or smeared moduli of the walled laminate of the tube

is used. The stiffness of the tube is then calculated using the conventional formula that is

commonly used for isotropic tubes as shown below.

Axial Stiffness ( )2 2x x o iE A E R Rπ= ⋅ − (2.11)

Bending Stiffness ( )4 4

4x x o iE I E R Rπ= ⋅ −

xE is the equivalent modulus of the tube in the direction. x

2.6 Bending Stiffness Comparison between the Approaches

In Ref [2], Chan and Demirhan calculated the bending stiffness of a composite

tube and compared their results with the smear property approach as well as the finite

element method, The model is also confirmed by experimental results [19].

15

CHAPTER 3

TAPERED TUBES UNDER AXIAL TENSION

3.1 Analytical Solution for an Isotropic Tube

Consider a uniform thickness tapered tube with a mid-thickness larger radius LR

and a mid-thickness smaller radius , where t is the thickness of the tube as shown in

Fig 3.1. Let

SR

L be the length of the tube.

Fig 3.1 Dimensions of the Tapered Tube

The area of the larger end of the tapered tube will be:

2 2

2

2

L L L

L L

t tA R R

A R t

π

π

2⎡ ⎤⎛ ⎞ ⎛ ⎞= + − −⎢ ⎥⎜ ⎟ ⎜ ⎟⎝ ⎠ ⎝ ⎠⎢ ⎥⎣ ⎦

=

(3.1)

16

Similarly, the area at the smaller end of the tube will be given by:

2S SA R tπ= (3.2)

The mid-thickness radius of the tapered tube at a distance x from the larger end, xR will

be given by:

tanx LR R x α= − (3.3)

Where α is the taper angle of the tube and Rtan

Lα ∆= ; R∆ is the difference between

the large and small radii of the tube;

L SR R R∆ = −

Hence, the area of the section of the tube at a distance x is given by:

22 [ tan

x x

x L

A R tA t R x ]

ππ α

== −

(3.4)

We know, the deformation of the tapered tube at a distance x , subjected to an axial force

F will be,

( )

( )

0

0

0

2 ta

2 t

x

xx

x

L

x

L

FdxEA

FdxEt R x

F dxEt R x

δ

n

an

π α

π α

= ∫

= ∫−

= ∫−

(3.5)

The axial deformation of the tapered bar xδ at any section, distance x from the fixed end,

will be hence given by:

17

( )

ln2 tan tan

Lx

L

RFE t R x

δπ α α

⎛ ⎞= ⎜⎜ −⎝ ⎠

⎟⎟ (3.6)

For a limiting case of a cylindrical tube, where 0α = the deformation can be found as:

( )0 0

lim lim ln2 tan tan

Lx

L

RFE t R xα α

δπ α α→ →

⎛ ⎞⎛ ⎞= ⎜ ⎜⎜⎜ −⎝ ⎠⎝ ⎠

⎟⎟⎟⎟ (3.7)

Using L’ Hospitale's rule, we can find out for an axial case,

2x

L

FxE t R

δπ

= (3.8)

The above equation can be verified from a solid mechanics text book.

3.2 Finite Element Model of an Isotropic Tube

3.2.1 Geometry and Modeling

A model of a tapered tube with dimensions LIR = 0.75 in, SIR = 0.25 in, and a

total thickness t (6 plyt ) = 0.03 in is created using ANSYS. LIR is the inner larger radius

and SIR is the inner smaller radius of the tube.

Refer to Fig 3.2. Key points 11 (0, )LIR , 12 ( , 21 ( ,, )SIL R )S plyL R t+ ,

22 (0, )L plyR t+ , are created using K command on ANSYS. Areas are created using these

key points using the A command. In a similar fashion, 6 areas are created by

incrementing the y co-ordinate by plyt . These areas are then swept about a longitudinal

18

axis to create volumes using the VROTAT command. The VROTAT command

necessitates that at least 2 volumes be created along the circumference. Thus a model of

the tube, consisting of 12 volumes is created.

AL 2024/T3 with properties, E = 1.0498e7 and ν = 0.33 is used as the material of the

tube.

Fig 3.2 Development of the Finite Element Model

3.2.2 Development of the Model

To establish the correct mesh size, the model has to be checked for convergence

of results. The convergence test for the tapered tube is carried out as follows. A model of

an isotropic (Al 2024/T3) tapered tube with a geometry discussed in Section 3.2.1 is

created. The element, SOLID 95, which has 20 nodes with 3 translational degrees of

freedom ( andUZ ) at each node is used for meshing the model. The mesh size is

varied from 720 to 9000 elements and an axial force of 500 lb is applied on the tube. The

mesh size of the lines in the thickness direction is maintained constant, and is equal to 1

element per line. The mesh sizes of the lines along the length and along the

,UX UY

19

circumference are varied. The axial deformation UX of the tube is noted for each mesh

size and are tabulated in Table 3.1

Table 3.1 Table showing the convergence of results with increase in the number of elements

No. of Elements along Thickness

No. of Elements along Length

No. of Elements along

Circumference

Total No. of elements

UX from Finite Element Method

(in) 6 10 12 720 2.990E-03

6 20 20 2400 3.002E-03

6 30 24 4320 3.007E-03

6 40 24 5760 3.010E-03

6 50 30 9000 3.010E-03

Fig 3.3 shows the graph of axial deformation versus the number of elements in

the model. From the plot, it can be seen that the results start converging at 5760 elements.

A mesh size of 9000 elements with 50 elements per length, 30 per circumference is

adopted in development of the model.

3.2.3 Boundary Conditions:

Fig 3.4 shows a model of the tapered tube with the boundary conditions. A

MASS21 element is created at the master node (2) located at the center of the smaller end

of the tube. Using CERIG command, a rigid region is created at the smaller end of the

tube in order to ensure uniform displacement in a section. At the larger end, all the nodes

are fixed in the ( ) direction. A force of 500 is applied on the master node in the

axial ( ) direction.

x FX lb

x

20

1.066E-05

1.068E-05

1.070E-05

1.072E-05

1.074E-05

1.076E-05

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

Number of Elements

Axi

al d

ispl

acem

ent '

UX

' in/

lb

Normalized UX fromFEM

Fig 3.3 Convergence Plot

Fig 3.4 Finite Element Model of an Isotopic Tube with Boundary Conditions

21

3.3 Results Comparison

Table 3.2 shows the comparison of the results between the analytic solution and the

finite element method. The percentage difference was calculated using the percent value

of the ratio of the difference between the analytical and FEM results to the FEM value.

Fig 3.5 shows the graph of displacement of the tube in direction, ‘UX ’ with respect to

the distance from fixed end ‘ ’. It can be observed from the graph that the analytic

solution is in excellent agreement with the finite element results. Thus the finite element

model developed can make a good ground for comparison of the analytical model and

hence can be used for modeling the composite tube.

x

x

Table 3.2 Comparison of normalized axial deformation from analytical method and finite

element method Distance from fixed

end x ( )in

Normalized Analytical

Solution ( / )in lb

Normalized FEM

Solution ( / )in lb%

Difference

1 6.83E-07 7.00E-07 2.58

2 1.41E-06 1.44E-06 1.55

3 2.20E-06 2.23E-06 1.17

4 3.06E-06 3.09E-06 0.95

5 3.99E-06 4.03E-06 0.85

6 5.03E-06 5.07E-06 0.78

7 6.18E-06 6.23E-06 0.72

8 7.47E-06 7.53E-06 0.67

9 8.96E-06 9.02E-06 0.65

10 1.07E-05 1.08E-05 0.43

22

0.00E+00

2.00E-06

4.00E-06

6.00E-06

8.00E-06

1.00E-05

1.20E-05

0 2 4 6 8 10

Distance from fixed end 'x'(in)

Nor

mal

ized

axi

al d

efor

mat

ion

'UX

' (in

/lb)

12

Results from analytic solution

Results from FEM Solution

Taper angle, α=2.86 Deg

Fig 3.5 Graph of Normalized Displacement Vs Distance from Fixed End

3.4 Analytical Model of the Composite Tube

3.4.1 Geometry

Consider a tapered tube with mid-thickness larger radius LR , mid-thickness smaller

radius SR , Length L and a tube thickness . Consider a section of the tapered tube at a

distance ‘

t

x ’ from the larger end, having a mid-thickness radius, xR . Note that

as shown in Fig 3.6.

6 plyt t=

3.4.2 Stiffness Matrices

The composite tapered tube as we know does not have a uniform radius, but the

radius varies linearly with the distance. Hence the stiffness matrices for the tube given in

23

equation 3.13 have to be written in terms of xR , where xR is the mid-thickness radius of

the tube at any section.

2'

0

2'

0

2'

0

x

x

x

A R A d

B R B d

D R D d

π

π

π

θ

θ

θ

⎡ ⎤⎡ ⎤ =⎣ ⎦ ⎣ ⎦

⎡ ⎤⎡ ⎤ =⎣ ⎦ ⎣ ⎦

⎡ ⎤⎡ ⎤ =⎣ ⎦ ⎣ ⎦

∫

∫

∫

(3.9)

Fig 3.6 Geometry of the Tapered Composite Tube

The individual terms of the ,A B⎡ ⎤ ⎡ ⎤⎣ ⎦ ⎣ ⎦ and D⎡ ⎤⎣ ⎦ are given in Appendix A.

24

Note that the reduced stiffness matrices ijQ of the plies are not only dependent on the

fiber orientation angle β and the elastic constants of the composite material, but are also

dependent on the orientation angle θ of the infinitesimal element with the axis. The

orientation angle term,

z

θ , however vanishes during the integration over the θ domain.

Fig 3.7 shows a flow chart of the complete procedure of finding the analytical model for

axial and torsion deformation for a tapered laminated composite tube. Note that the order

of the transformation has no effect on the final stiffness matrices of the tube.

Fig 3.7 Development of the Analytical Model from the Laminated Plate Theory

25

3.4.3 Axial stiffness and Stress Distribution

From equation 2.8, we saw that the axial deformation for a uniform laminated

composite tube is given by

110x

xa Nε = (3.10)

Where, is the total force in the direction. xN x

For a tapered tube, the axial deformation has to be integrated with respect to

since

x

11a is a function of . The total axial deformation is thus given by: x

Fig 3.8 Flow Chart of the Development of the Analytical model

26

110

x

x F a dXδ = ∫ (3.11)

Note that the deformation remains the same at all points in a cross sectional plane.

From the equation for as given in Appendix A, we can see that the term

has a unit lb . Hence is of the unit,

11A 11A

11a 1lb

. The product of force and which is

dimensionless is integrated with respect to . Thus we can see that the equation on the

right has a dimension of .

F 11a

x

in

For a tube of uniform wall thickness, the B⎡ ⎤⎣ ⎦ matrix will be zero, irrespective of

whether the lay-up is symmetric or asymmetric implying that there is no extension-

bending coupling in a laminated tube. Hence, 1

a A−

⎡ ⎤ ⎡ ⎤=⎣ ⎦ ⎣ ⎦ . Since, A⎡ ⎤⎣ ⎦ is a function of

the laminate constants and xR , xR can be taken out and the laminated constants can be

substituted to give 11a in the form of 1

xR multiplied by constant.

111

x

a CR

= ⋅

The constant can be found out by substituting the elastic constants in the C A terms and

then obtaining the first row, first column term of the matrix, a⎡ ⎤⎣ ⎦ . Thus the axial

deformation of the tube can be written as:

27

( )

lntan tan

Lx

L

RFCR x

δα α

⎛ ⎞= ⎜⎜ −⎝ ⎠

⎟⎟ (3.12)

Note that the term is a function of11a xR , which is given by:

tanx LR R x α= −

The equation 3.11 is analogous to the equation for an isotropic tube given in Section 3.1,

where is replaced by 11a 1

xE A. Note that the term takes into account the material

properties and the effect of curvature.

11a

3.4.4 Calculation of the Stresses from Analytical Solution

The analytical model developed is further validated for stresses in the layers. The

stresses are computed as follows.

We know that,

0T

a b NMb d

εκ

⎡ ⎤⎡ ⎤ ⎡= ⋅⎢ ⎥

⎤⎢ ⎥

⎢ ⎥⎢ ⎥

⎣ ⎦ ⎣⎣ ⎦ ⎦ (3.13)

.

For a laminate tube with uniform thickness, 0b⎡ ⎤ =⎣ ⎦ . For the tension case, only will

be applied. Hence, [

xN

] 0κ = and [ ] [ ]0 a Nε ⎡ ⎤= ⋅⎣ ⎦ . Hence, for any section, stresses in the

global co-ordinate system are given by:

28

[ ] [ ]1 2 x yQ aσ

− −⎡ ⎤⎡ ⎤= ⋅ ⋅⎣ ⎦ ⎣ ⎦ N

The a⎡ ⎤⎣ ⎦ matrix is a function of xR and can be found out for different sections by

changing from 0 to x L .

3.5 Finite Element Model of the Composite Tapered Tube

3.5.1 Geometry of the tube

A tube made of 6 / 3501 6IM − Graphite Epoxy composite with a lay-up [ ]45/ 0S

±

is considered. The dimensions of the tube are same as the isotropic case as described in

Section 3.2.1. The ply thickness = 0.005 in plyt

The composite material 6 / 3501 6IM − has the following properties.

1 2 3

12 23 13

12 23 13

22.8 6 ; 1.35 6 ; 1.35 6

0.3; 0.34; 0.3;

0.83 6 ; 0.504 6 ; 0.83 6

E E psi E E psi E E psi

G E psi G E psi G E psi

ν ν ν

= = =

= = =

= = =

3.5.2 Meshing the model

3.5.2.1 Element Description

The element used for meshing is SOLID 191, which is the layered version of

SOLID95 element. The element has 20 nodes, with and UZ degrees of freedom ,UX UY

29

at each node. The element allows up to 100 layers. If more layers have to be input,

elements have to be stacked. Fig 3.9 shows the geometry of the Solid 191 Element.

The material properties can be input in two ways. The first method is to input the

material constants and defining the lay-up of the laminate. The second method is to input

the laminate constitutive matrix, which is computed by an external program.

While creating the model, care has to be taken in order to insure that the co-ordinate is

oriented along the thickness.

z

Fig 3.9 Solid 191 Element Geometry

3.5.2.2 Generation of the Model

6 volumes, corresponding to 6 plies are created in the model. A real constant set is

defined for each ply and assigned to the corresponding volume using the VAT

command. The volumes are naturally glued to each other because of the method of

T

30

creation of volumes. The mesh sizes, obtained from the convergence test are specified in

order to ensure that correct results are obtained from finite element method.

3.5.3 Boundary Conditions

The model is constrained at the larger (left) end for UX only. A rigid region is

created at the smaller end using the CERIG command. A force of 500 lb is applied

on the master node.

FX

3.6 Results Comparison

Since a rigid region is created, all nodes in a section have the same displacement.

The values of UX, which is the deformation in direction at 11 different sections are

tabulated and compared with the analytical solution. Table 3.3 gives the comparison

between the analytic solution results and the finite element results. Fig 3.10 shows the

graph of normalized axial deformation versus distance from fixed end

x

Fig 3.11 shows the axial deformation of the tapered tube at its tip. It is shown that

unlike a uniform tube, non linear deformation along the length direction is obtained.

31

Table 3.3 Comparison of normalized axial deformation from analytical method and finite element method for a composite tube

Distance from Fixed End (in)

Normalized UX- FEM (in/lb)

Normalized UX- Analytic (in/lb)

% Difference

1 7.94E-07 8.14E-07 2.52

2 1.60E-06 1.71E-06 6.88

3 2.47E-06 2.65E-06 7.29

4 3.41E-06 3.68E-06 7.80

5 4.45E-06 4.80E-06 7.98

6 5.58E-06 6.06E-06 8.57

7 6.85E-06 7.44E-06 8.65

8 8.28E-06 9.00E-06 8.75

9 9.92E-06 1.08E-05 8.91

10 1.18E-05 1.29E-05 9.05

0.00E+00

2.00E-06

4.00E-06

6.00E-06

8.00E-06

1.00E-05

1.20E-05

1.40E-05

0 2 4 6 8 10 1

Distance from fixed end 'x' (in)

Nor

mal

ized

axi

al d

efor

mat

ion

'UX

' (in

/lb

2

) Results from FEM

Results from analytical

Taper angle, α=2.86

Fig 3.10 Graph of axial deformation vs. distance from fixed end for the composite tube

32

Fig 3.11 Plot showing the displacement of the tip of the tapered tube

Fig 3.12 Stresses along the direction in the ply x 045+

33

Fig 3.13 Stresses along the direction in the ply x 045−

Fig 3.14 Stresses along the direction in the ply x 00

34

Figs. 3.12, 3.13 and 3.14 show the stresses along the direction in ,

and plies starting from the innermost layer. It can be seen that the maximum

stresses occur at the smallest cross section. This is however not the tip of the tube, since

after deformation, the area of the tip of the tube is enlarged. It can be observed that the

maximum stresses of the and the plies are close; Comparing figures 3.12, 3.13

and 3.14 we can find that the layer has the highest maximum stresses.

x 045+

045− 00

045+ 045−

00

3.6.1 Comparison of Axial Stresses

Stresses are calculated in the layer using the analytical model developed in

Section 3.5.1 and are compared with the finite element results. The results are tabulated

in Table 3.4

045+

Table 3.4 Comparison of stresses

Stresses Analytical solution (psi) Finite element method (psi)

x =5 1463 1468.5

x =10 3155.3 3187.9

35

CHAPTER 4

TAPERED TUBES UNDER TORSION

4.1 Analytical Solution for an Isotropic Tube

Consider a tapered tube with geometry as mentioned in Section 3.1. The radius of

the tube at a distance from the fixed end x xR is given by:

tanx LR R x α= − ,

Where α the taper is angle of the tube and tan RL

α ∆= ; R∆ is the difference between the

radii of the tube;

L SR R R∆ = −

The polar moment of inertia at the large section of the tube will be given by: LJ

4 4

2 2 2L L LtJ R Rπ t⎡ ⎤⎛ ⎞ ⎛ ⎞= + − −⎢ ⎥⎜ ⎟ ⎜ ⎟

⎝ ⎠ ⎝ ⎠⎢ ⎥⎣ ⎦ (4.1)

It can be simplified to form

342L L LJ R t R 3tπ ⎡ ⎤= +⎣ ⎦ (4.2)

Similarly the polar moment of inertia of the tube at a section from the fixed end is

given by

x

36

342x x x

3J R t R tπ ⎡ ⎤= +⎣ ⎦ (4.3)

The angle of twist of a bar with varying cross section at a section, distance

from the fixed end is given by:

x

0

x

xx

T dXG J

φ = ∫ (4.4)

After integration, the angle of twist of the bar at any section is given by:

( )( )

( )( )

22 2

3 2 2

4 tanln

tan 4 tan

L Lx

L L

R t R xTG t t R R x

αφ

π α α

⎛ ⎞+ −⎜= ⋅ ⎜ + −⎜ ⎟⎝ ⎠

2⎟⎟ (4.5)

For the limiting case of a cylinder, tan 0α = . Hence using L’Hospitale’s rule, we can

find the twisting angle as:

( )3 34

2

Tx

G R t Rtφ π=

⋅ +

This can be verified from a solid mechanics text book.

A MATHEMATICA program (Appendix C) is used to calculate the angle of twist

of a tapered tube with the following dimensions: LR = 0.765 in; SR = 0.265 in; L = 10 in;

= 0.03 in; t

The results are compared with finite element results and tabulated in Table 4.1.

37

4.2 Finite Element Model of an Isotropic Tube

A tapered tube of dimensions, LR = 0.765 in ; SR = 0.265 in, L = 10 in and t =

0.03 in is modeled in the same fashion as described in Section 3.2. Appendix (E) shows

the ANSYS batch file used for the generation of the model. Fig 4.1 shows the finite

element model of the tapered tube along with the boundary conditions and loads.

4.2.1 Meshing the Model

The same model as described in Section 3.7 is adopted. The model is fixed at the

large end for all degree of freedom as shown in Fig 4.1. At the smaller end, a rigid region

is created using CERIG command. A torque (MX) of 280 lb-in is applied at the master

node.

Fig 4.1 Finite Element Model of a Tapered Isotropic Tube under Torsion with Boundary Conditions

38

4.3 Results Comparison

Table 4.1 shows the comparison of the results obtained from FEM and Analytical

solution. Fig 4.2 shows a graph of normalized angle of twist results versus distance from

the fixed end. It can be observed that the finite element model results are very close to the

analytical results with a low percentage error of 0.6.

Table 4.1 Comparison of angle of twist results from analytical method and finite element method for an isotropic tube

Distance from the fixed End ( )x in

Normalized angle of twist- FEM ( / )in lb

Normalized angle of twist -Analytic ( / )in lb

% Difference

1 3.32E-06 3.32E-06 0.00

2 7.43E-06 7.42E-06 0.13

3 1.25E-05 1.25E-05 0.00

4 1.91E-05 1.91E-05 0.00

5 2.77E-05 2.76E-05 0.36

6 3.92E-05 3.91E-05 0.26

7 5.51E-05 5.50E-05 0.18

8 7.80E-05 7.78E-05 0.26

9 1.13E-04 1.12E-04 0.89

10 1.69E-04 1.68E-04 0.60

39

0.00E+00

1.00E-02

2.00E-02

3.00E-02

4.00E-02

5.00E-02

0 2 4 6 8 10Distance from fixed end 'X'

Nor

mal

ized

ang

le o

f tw

ist 'p

hi/T

'

12

FEM SolutionAnalytical solution

Taper angle, α=2.86

Fig 4.2 Graph showing the normalized angle of twist from analytical and finite element method Vs distance from fixed end

4.4 Analytical Model of the Composite Tube

In Section 3.3, we derived the laminate stiffness equations of the tube

,A B⎡ ⎤ ⎡ ⎤⎣ ⎦ ⎣ ⎦ and D⎡ ⎤⎣ ⎦ at any section of radius xR (2.7). We know that the mid-plane strains

and curvatures of a laminate are given by:

0

T

a b N

Mk b d

ε ⎡ ⎤⎧ ⎫ ⎧ ⎫⎪ ⎪ ⎪⎢ ⎥= ⎪⎨ ⎬ ⎨

⎢ ⎥⎬

⎪ ⎪⎪ ⎪ ⎩ ⎭⎩ ⎭ ⎣ ⎦ (4.6)

where,

1

T

a b A B

B Db d

−⎡ ⎤ ⎡ ⎤⎢ ⎥ ⎢ ⎥=⎢ ⎥ ⎢ ⎥⎣ ⎦⎣ ⎦

Now, expanding the above matrix for xyκ , we have

16 26 61 62 6666xy x y xy x y xyk b N b N b N d M d M d M= + + + + + (4.7)

40

Since in the problem, we are applying only a torque T ( xyM in composite terminology),

all the terms except 66 xyd M are reduced to zero. Hence,

66xy xyd Mκ = (4.8)

For the tapered tube, the above equation has to be integrated with respect to , since the

area of the tube varies linearly with respect to . Hence, the angle of twist at any section,

x

x

xφ ( in composite terminology) is given by: xyκ

66

0

x

x T d dxφ = ∫ (4.9)

Where andxyT M= x xyφ κ= .

Since 0B⎡ ⎤ =⎣ ⎦ for a tube of uniform thickness,

1d D

−⎡ ⎤ ⎡ ⎤=⎣ ⎦ ⎣ ⎦

All the constants are plugged into the D matrix terms and then an inverse of the

matrix is computed. The 66d term is then a function of 3xR and xR . This is further used in

the integration to compute the angle of twist at any section. Note that in case of a

laminate, xyM has a unit of inlbin

− , but in case of the tapered tube, T or xyM has units

. Note that term lb in− 66d has a unit of, 2

1lb in−

.Thus the angle of twist has a unit of

. rad

41

4.4.1 Stress Calculations

For a tube of uniform laminate, 0B⎡ ⎤ =⎣ ⎦ . Hence, from equation 3.13,

[ ] [ ]d Mκ ⎡ ⎤= ⋅⎣ ⎦ . In the torsion case only xyM is present. Hence, the stresses, can be found

as:

[ ] [ ]kkx y k

Q z d Mσ−

⎡ ⎤ ⎡ ⎤= ⋅ ⋅ ⋅⎣ ⎦ ⎣ ⎦ (4.10)

Using this method, stresses are calculated for the plies. Note that in the above equation,

is the distance of the surface from the mid-section of the tube. kz

4.5 Finite Element Model of the Composite tube

The tapered tube model generated in Section 3.7 is used in the torsion case as

well. The boundary conditions applied are however different in this case. The tube is

fixed for all DOF at the larger end as shown in Fig 4.1. At the smaller end, a rigid region

is created using command. A torque CERIG 280T lb in= − is applied on the master node.

The model is then solved for its stresses and strains.

42

Fig 4.3 Calculation of the Angle of twist from Finite Element Model

4.5.1 Post Processing and Results Comparison

The angle of twist φ can be obtained from the post processor as nodal solution

for . However, these results are not calculated at every node. Hence a manual

method for calculation of

ROTX

φ is adopted.

Refer to Fig 4.3. Under the application of a torque, the point on the circumference moves

from A to A′ . The angle swept is the angle of twistφ and is the arc length. Now, we

know that the arc length is given by,

S

S R φ= , where R is the radius. The arc length will

be approximately equal to the hypotenuse

S

AA′ . Hence, 2~S UY UZ+ 2 . Thus angle of

twist, 2 2

~UY UZ

Rφ

+. and are the displacements in Y and UY UZ Z respectively,

obtained from the post processor.

43

Table 4.2 shows the results comparison between analytic solution and Finite

Element Method for the angle of twist of a laminated tapered tube. Fig 4.4 shows the

comparison of the angle of twist solution between analytical and finite element methods

for different sections of the tube.

Table 4.2 Comparison of angle of twist results from analytical method and finite element method for a composite tube

Distance from fixed

End (in) xNormalized φ from FEM

( /rad lb in− ) Normalized φ from

Analytic ( /rad lb in− ) %

Difference 1 3.771E-06 3.929E-06 4.16

2 8.312E-06 8.679E-06 4.41

3 1.404E-05 1.471E-05 4.78

4 2.137E-05 2.250E-05 5.29

5 3.092E-05 3.286E-05 6.23

6 4.372E-05 4.714E-05 7.83

7 6.140E-05 6.679E-05 8.76

8 8.680E-5 9.482E-05 9.24

9 1.252E-04 1.386E-04 9.28

10 1.865E-04 2.089E-04 11.63

44

0.000E+00

5.000E-05

1.000E-04

1.500E-04

2.000E-04

2.500E-04

0 2 4 6 8 10 12

Distance from the fixed end 'X' (in)

Nor

mal

ized

Ang

le o

f tw

ist ∅

(rad

/lb-in

)Results from analytic solution

Results from FEM

Taper angle, α=2.86 deg

Fig 4.4 Graph of normalized angle of twist vs. Distance from fixed End

Fig 4.5 Shear stresses on the ply 045+

45

Fig 4.6 Shear stresses on the ply 045−

Fig 4.7 Shear stresses on the ply 00

46

Fig.s 4.5, 4.6 and 4.7 show the shear stress plots of the , and plies. It can be

seen that the and the plies have nearly the same maximum stresses. Also, the

layer has lesser maximum stress than the and the layers. This can be

explained because the layer has lesser equivalent shear modulus; hence higher shear

deformation and hence lesser stresses.

045+ 045− 00

045+ 045−

00 045+ 045−

00

4.5.2 Comparison of Shear Stresses ( xyτ )

Stresses from the analytical method are found using 4.10. Stresses are evaluated

for the (innermost) layer at two different sections. Table 4.3 shows the comparison

of shear stresses between the analytical and FEM results.

045+

Table 4.3 Comparison of Shear stresses

Stresses Analytical solution (psi) Finite element method (psi)

x =5 5998 6041

x =10 22576 22684

It can be seen from the table that the results are in good agreement with each other.

47

CHAPTER 5

PARAMETRIC STUDIES

The geometry of the tube, stacking sequence, taper angle, fiber orientation etc.

play an important role in the stiffness of composite tubes. In this study, an effort has to be

done to study the effects of these on the deformations of the tube. IM6/3501-6 Graphite

Epoxy composite is used for these studies. A tube with dimensions as described in

Section 3.6.1 is considered for the analysis.

5.1 Stacking Sequence Effect

5.1.1 Axial Case

In this section, the effect of stacking sequence on the axial deformation of the

tube is studied. To examine the effects, three laminates having the same fiber

orientations, but different arrangement is considered. [ ]45/ 0S

± , [ ]0 / 45S

± and

[ ]45/ 0 / 45+ − are the three lay-ups considered. The variation of the analytical solution

for the three lay-ups is seen. MATHEMATICA program (Appendix D) is used for the

calculation of the results, which are tabulated in Table 5.1

48

Table 5.1 Variation of axial deformation with stacking sequence

Lay-Up Normalized Axial Deformation ‘ / Fδ ’ (in/lb)

[ ]45 / 0S

± 1.294E-4

[ ]0 / 45S

± 1.294E-4

[ ]45/ 0 / 45+ − 1.294E-4

It can be seen that the stacking sequence has no effect on the axial deformation of

the tapered tube. This can be explained, because the extensional stiffness matrix of the

tube A⎡ ⎤⎣ ⎦ does not vary with the arrangement of the plies.

5.1.2 Torsion Case

Here, the effects of stacking sequence on the torsion deformation (twisting angle)

are studied. The same lay-ups as in Section 5.1.1 are considered. The results are

generated using the MATHEMATICA code (Appendix D). Table 5.2 shows the angle of

twist results for the different lay-ups.

Table 5.2 Variation of angle of twist with stacking sequence

Lay-Up

Normalized Angle of twist /Tφ (rad/lb-in)

[ ]45 / 0S

± 3.2646E-04

[ ]0 / 45S

± 3.2676E-04

[ ]45/ 0 / 45+ − 3.2675E-04

49

It can be observed from the table that stacking sequence has little effect on the

angle of twist of the tapered tube. The twisting angle is dependent on the torsional

stiffness of the tube cross section. The torsional stiffness of the laminate 66

1d

does not

depend on the stacking sequence. However, the torsional stiffness of the tube is

predominant by the A⎡ ⎤⎣ ⎦ matrix, which is independent of stacking sequence. As a result,

the twisting angle has little effect on the stacking sequence of the laminate.

5.2 Effect of Fiber Orientation

5.2.1 Axial case

One of the important properties of laminated composite materials is their

directional dependence, which facilitates a design which can be tailored to the

application. This property is exploited in almost all the applications of laminated

composites. In the current study, symmetric lay-ups with θ+ o , θ− o and 0 plies are

considered.

o

θ values are changed from 15 to 75 with an increment of 15 . The

MATHEMATICA code (Appendix D) is used to calculate the results. Results are

tabulated in Table 5.3

o o o

50

Table 5.3 Variation of axial deformation with fiber orientation

Lay-Up Normalized Axial Deformation / Fδ (in/lb)

[ ]15 / 0S

± 5.888E-6

[ ]30 / 0S

± 9.982E-6

[ ]45 / 0S

± 1.295E-5

[ ]60 / 0S

± 1.383E-5

[ ]75 / 0S

± 1.405E-5

4.00E-06

6.00E-06

8.00E-06

1.00E-05

1.20E-05

1.40E-05

1.60E-05

0 10 20 30 40 50 60 70 80

Angle (Deg)

Nor

mal

ized

axi

al d

efor

mat

ion

(in/lb

)

Fig 5.1 Variation of axial deformation with fiber orientation

It can be observed from the table that the deformation is smallest for the 15

orientation and increases with the fiber orientation. This can be accounted for by the fact

that a 15 orientation is stiffer in the direction, since fiber direction and loading

directions are separated by a small angle. As the fiber orientation increases the fiber

direction becomes almost transverse to the loading direction and hence the stiffness in

loading direction ( ) becomes less.

o

o x

x

51

5.2.2 Torsion Case

The effects of fiber orientation are studied in this section. The same fiber

orientations as in Section 5.2.1 are considered. Normalized angle of twist are computed

using the MATHEMATICA code (Appendix D) for the different orientations. The results

are tabulated in Table 5.4

Table 5.4 Variation of angle of twist with fiber orientation

Lay-Up Normalized angle of twist /Tφ (rad/lb)

[ ]15 / 0S

± 1.017E-4

[ ]30 / 0S

± 4.709E-5

[ ]45 / 0S

± 3.712E-5

[ ]60 / 0S

± 4.709E-5

[ ]75 / 0S

± 1.017E-4

From Fig 5.2, we can observe that the angle of twist is the least for the orientation.

Also, the angle of twist is the same for complementary angle orientations ( );

For eg., 30 and 60 orientations have the same angle of twist.

45o

1 2 90θ θ+ = o

o o

This can be explained by the fact that the equivalent shear modulus is the

least for 45 orientation as shown in Fig. 5.3.

xyG

o

52

0.00E+00

2.00E-05

4.00E-05

6.00E-05

8.00E-05

1.00E-04

1.20E-04

0 10 20 30 40 50 60 70 80

Angle (Deg)

Nor

mal

ized

twis

ting

angl

e (r

ad/lb

-in)

Taper angle, α=2.86 Deg

Fig 5.2 Variation of angle of twist with fiber orientation

0.00E+00

1.00E+06

2.00E+06

3.00E+06

4.00E+06

5.00E+06

6.00E+06

7.00E+06

0 20 40 60 80

Angle (Deg)

Equ

ival

ent S

hear

mod

ulus

(G

xy)

Fig. 5.3 Equivalent shear modulus plot

53

5.3 Effect of the Taper Angle

5.3.1 Axial case

In this section, the effects of taper angle, α on the deformations of the tube are

studied. For a large taper angle, the fiber orientation in the ply does not remain uniform,

which results in erroneous results. The MATHEMATICA code given in Appendix D is

used for the calculation; the results are tabulated in Table 5.5. For the calculation, the

larger radius LR is kept constant and the smaller radius SR varied accordingly.

Table 5.5 Variation of axial deformation with taper angle α

Taper angle ( )rad Deg Normalized Axial Deformation ( ) /in lb

0.099 (5.67) 9.713E-06

0.162 (9.28) 1.107E-05

0.219 (12.55) 1.219E-05

0.275 (15.76) 1.315E-05

0.328 (18.79) 1.402E-05

0.381 (21.83) 1.476E-05

0.433 (24.81) 1.544E-05

0.485 (27.79) 1.606E-05

Fig 5.4 shows the graph of variation of deformation with radius ratio. It can be

seen that for tubes with low taper angles, the deformation of the bar is less. This is

because tubes with low taper angles are nearly cylindrical in nature and are stiffer than

tubes with high taper angle.

54

8.000E-06

1.000E-05

1.200E-05

1.400E-05

1.600E-05

1.800E-05

0.070 0.110 0.150 0.190 0.230 0.270 0.310 0.350 0.390 0.430 0.470

Taper angle (rad)

Nor

mal

ized

Axi

al d

efor

mat

ion

(in/lb

)

Fig 5.4 Variation of normalized axial deformation with taper angle

5.3.2 Torsion case

The variation of angle of twist for tapered tubes with different taper angles α is

studied here. For the calculations, the larger radius LR is kept constant and the smaller

radius SR varied accordingly. Taper angles varying from 0.099 to 0.485 are used to

compute the normalized angle of twist. The results are tabulated in Table 5.6.

From the Fig 5.5, we can observe that the twisting angle increases with the

increase in the ratio. Low taper angle tubes are stiffer than high taper angle tubes and

hence have lesser deformation.

55

Table 5.6 Variation of angle of twist with taper angle α

Taper angle ( )rad Deg Normalized Angle of twist /Tφ (rad/lb)

0.099 (5.67) 1.092E-04

0.162 (9.28) 1.748E-04

0.219 (12.55) 2.548E-04

0.275 (15.76) 3.494E-04

0.328 (18.79) 4.607E-04

0.381 (21.83) 5.822E-04

0.433 (24.81) 7.204E-04

0.485 (27.79) 8.730E-04

0.000E+00

2.000E-04

4.000E-04

6.000E-04

8.000E-04

1.000E-03

0.070 0.110 0.150 0.190 0.230 0.270 0.310 0.350 0.390 0.430 0.470

Taper angle (rad)

Nor

mal

ized

ang

le o

f tw

ist(r

ad/in

-lb)

Fig 5.5 Variation of normalized angle of twist with taper angle

56

CHAPTER 6

CONCLUSIONS

An analytical model for the deformation of tapered laminated composite tubes

under axial tension and torsion was developed in this research. The derivation was based

on the Laminated Plate approach. The results from analytical model are compared with

the results from finite element method. From this research, the following conclusions can

be made.

The axial deformation and the twisting angle for composite tube under axial load

and twisting moment respectively predicted by the current developed method

agree well with the results obtained from finite element analysis.

The axial stresses and the shear stress for each ply are also in excellent agreement

with the results obtained from the finite element method.

Angle of twist can be minimized by incorporating 45o plies.

Axial deformation of composite tubes can be minimized by incorporating near 0o

plies.

Given the set of plies, they can be stacked in any order since the sequence does

not affect the deformations significantly.

Tubes with less taper angle can be employed to minimize the deformations in

structural applications.

57

APPENDIX A

TERMS OF THE STIFFNESS MATRICES

58

( ) ( )

( ) ( )

: cos ; sin ; . .

34 2 2 42 211 11 12 66 22 11 8

3 12 2 4 42 212 11 22 66 12 11 8 2

4 2 22 222 11 12 6

Note c s Transformation terms w r to zz z

nA R c Q s c Q Q s Q z zz z z zX k kk k

nA R s c Q Q Q s c Q z zz z z zX k kk k

A R s Q s c Q Qz z zX

θ θ

π

π

π

= =

= + + + ⋅ −∑ −=

= + − + + ⋅ −∑ −=

= + +

⎡ ⎤⎢ ⎥⎣ ⎦

⎡ ⎤⎛ ⎞⎜ ⎟⎢ ⎥⎝ ⎠⎣ ⎦

( ) ( )

( )

( )

3 46 22 11 8

1 1 33 3216 11 12 66 12 22 66 11 2 2 8

1 1 33 3226 11 12 66 12 22 66 11 2 2 8

2266

nc Q z zz k kk k

nA R s c Q Q Q s c Q Q Q z zz z z zX k kk k

nA R c s Q Q Q c s Q Q Q z zz z z zX k kk k

A R s czX

π

π

π

+ ⋅ −∑ −=

= − − + − + ⋅∑ −=

= − − + − + ⋅∑ −=

=

⎡ ⎤⎢ ⎥⎣ ⎦

⎡ ⎤⎛ ⎞ ⎛ ⎞⎜ ⎟ ⎜ ⎟⎢ ⎥⎝ ⎠ ⎝ ⎠⎣ ⎦

⎡ ⎤⎛ ⎞ ⎛ ⎞⎜ ⎟ ⎜ ⎟⎢ ⎥⎝ ⎠ ⎝ ⎠⎣ ⎦

−

−

( ) ( )

( ) ( )

( ) ( )

3 12 4 411 22 12 66 66 11 8 2

34 2 2 4 2 2211 11 12 66 22 11 8

3 12 2 4 4 2 2212 11 22 66 66 11 8 2

422 11

nQ Q Q Q s c Q z zz z z k kk k

nB R c Q s c Q Q s Q z zz z z zX k kk k

nB R s c Q Q Q s c Q z zz z z zX k kk k

B R s Q szX

π

π

π

+ − − + + ⋅ −∑ −=

= + + + ⋅ −∑ −=

= + − + + ⋅ −∑ −=

= +

⎡ ⎤⎛ ⎞⎜ ⎟⎢ ⎥⎝ ⎠⎣ ⎦

⎡ ⎤⎢ ⎥⎣ ⎦

⎡ ⎤⎛ ⎞⎜ ⎟⎢ ⎥⎝ ⎠⎣ ⎦

( ) ( )

( )

32 2 4 2 2212 66 66 11 8

1 1 33 316 11 12 66 12 22 66 11 2 2 8

1 1 33 3226 11 12 66 12 22 661 2 2 8

nc Q Q c Q z zz z z k kk k

nB R s c Q Q Q s c Q Q Q z zz z z zX k kk k

nB R c s Q Q Q c s Q Q Q zz z z zX kk k

π

π

+ + ⋅ −∑ −=

= − − + − + ⋅∑ −=

= − − + − + ⋅∑=

⎡ ⎤⎢ ⎥⎣ ⎦

⎡ ⎤⎛ ⎞ ⎛ ⎞⎜ ⎟ ⎜ ⎟⎢ ⎥⎝ ⎠ ⎝ ⎠⎣ ⎦

⎡ ⎤⎛ ⎞ ⎛ ⎞⎜ ⎟ ⎜ ⎟⎢ ⎥⎝ ⎠ ⎝ ⎠⎣ ⎦

( )

2 2

2

−

−

( ) ( )

( ) ( )

( ) ( )

21

3 12 2 4 4 2 2266 11 22 12 66 66 11 8 2

2 34 2 2 4 3 3211 11 12 66 22 113 8

1 3 53 4 2 2 42 211 12 66 22 11 2 4 16

zk

nB R s c Q Q Q Q s c Q z zz z z zX k kk k

R nXD c Q s c Q Q s Q z zz z z z k kk k

nR c Q s c Q Q s Q z zz z z zX k kk k

D

π

π

π

−

= + − − + + ⋅ −∑ −=

= + + + ⋅ −∑ −=

+ + + + ⋅ −∑ −=

⎡ ⎤⎛ ⎞⎜ ⎟⎢ ⎥⎝ ⎠⎣ ⎦

⎡ ⎤⎢ ⎥⎣ ⎦

⎡ ⎤⎢ ⎥⎣ ⎦

( ) ( )

( ) ( )

2 3 12 2 4 4 3 3212 11 22 66 66 113 8 2

1 5 3 33 2 2 4 42 11 22 66 12 11 2 16 2 8

R nX s c Q Q Q s c Q z zz z z z k kk k

nR s c Q Q Q s c Q z zz z z zX k kk k

π

π

= + − + + ⋅ −∑ −=

+ + − + + ⋅ −∑ −=

⎡ ⎤⎛ ⎞⎜ ⎟⎢ ⎥⎝ ⎠⎣ ⎦

⎡ ⎤⎛ ⎞⎜ ⎟⎢ ⎥⎝ ⎠⎣ ⎦

59

( ) ( )

( ) ( )

2 34 2 2 4 3 3222 11 12 66 22 113 8

1 3 53 4 2 2 42 211 12 66 22 11 2 4 16

2 1 1 33 3 316 11 12 66 12 22 6613 2 2 8

R nXD s Q s c Q Q c Q z zz z z z k kk k

nR s Q s c Q Q c Q z zz z z zX k kk k

R nXD s c Q Q Q s c Q Q Q z zz z z z kk k

π

π

π

= + + + ⋅ −∑ −=

+ + + + ⋅ −∑ −=

= − − + − + ⋅ −∑=

⎡ ⎤⎢ ⎥⎣ ⎦

⎡ ⎤⎢ ⎥⎣ ⎦

⎡ ⎤⎛ ⎞ ⎛ ⎞⎜ ⎟ ⎜ ⎟⎢ ⎥⎝ ⎠ ⎝ ⎠⎣ ⎦

( )

( )

( )

31

1 3 3 3 5 33 3 32 11 12 66 12 22 66 11 2 8 4 8 16 4

2 1 1 33 3 326 11 12 66 12 22 66 113 2 2 8

1 3 33 32 11 12 662 8 4

k

nR s c Q Q Q s c Q Q Q z zz z z zX k kk k

R nXD s c Q Q Q s c Q Q Q z zz z z z k kk k

R s c Q Q Qz zX

π

π

π

−

+ − − + − + ⋅∑

3

− −=

= − − + − + ⋅ −∑ −=

+ − −

⎡ ⎤⎛ ⎞ ⎛ ⎞⎜ ⎟ ⎜ ⎟⎢ ⎥⎝ ⎠ ⎝ ⎠⎣ ⎦

⎡ ⎤⎛ ⎞ ⎛ ⎞⎜ ⎟ ⎜ ⎟⎢ ⎥⎝ ⎠ ⎝ ⎠⎣ ⎦

⎛⎝

( )

( ) ( )

( )

3 5 3312 22 66 11 8 16 4

2 3 12 2 4 4 3 366 11 22 12 66 66 113 8 2

1 5 3 3 33 2 2 4 42 11 22 12 66 661 2 16 4 4 8

ns c Q Q Q z zz z k kk k

R nXD s c Q Q Q Q s c Q z zz z z z k kk k

nR s c Q Q Q Q s c Q z zz z z zX k kk k

π

π

+ − + ⋅ −∑ −=

= + − − + + ⋅ −∑ −=

+ + − + + ⋅ −∑ −=

⎡ ⎤⎞ ⎛ ⎞⎜ ⎟ ⎜ ⎟⎢ ⎥⎠ ⎝ ⎠⎣ ⎦

⎡ ⎤⎛ ⎞⎜ ⎟⎢ ⎥⎝ ⎠⎣ ⎦

⎡ ⎤⎛ ⎞⎜ ⎟⎢ ⎥⎝ ⎠⎣ ⎦

( )1

60

APPENDIX B

MATHEMATICA CODE FOR THE ANALYTICAL SOLUTION OF AXIAL

DEFORMATION OF AN ISOTROPIC TUBE

61

Off@General ::spell D

tanalpha =RL − Rs

L;

Rx = Rs + HL − xL tanalpha ;

Ax = π JJRx +t2N

2−JRx −

t2N

2N;

Def = J FY

‡x

AxN

−

F L Log@2 π t HHL− xL RL + x RsLD2 π t Y RL − RsH L

v1 = Def ê. 8RL → 0.765 , Rs → 0.265 , L → 10, t → 0.03 , F → 500, Y → 1.0498 107<;

v2 = Table B Hv1 ê. x → iL−Hv1 ê. x → 0L500

, 8i, 10<F êê ColumnForm

62

APPENDIX C

MATHEMATICA CODE FOR THE ANALYTICAL SOLUTION OF ANGLE OF TWIST OF AN ISOTROPIC TUBE

63

t = 0.030; Rs = 0.265; R = 0.765;

L = 10;

tanaplha =R− Rs

L;

T = 280; Y = 10.498106; ν = 0.33;

G =Y

2H1+nL ;

H*Rx=R- x tanaplha;*L

φ =T

π Gt3 tanaplhaLogA R2 Ht2 + 4HR− xtanaplhaL2L

Ht2 + 4 R2L HR− xtanaplhaL2E ê.x → 10

64

APPENDIX D

MATHEMATICA CODE FOR THE ANALYTICAL SOLUTION OF AXIAL DEFORMATION AND ANGLE OF TWIST OF A TAPERED

COMPOSITE TUBE

65

H∗ Closed form solution for stiffness of composite tube∗LOff@General::spellDE1= 22.8106; E2= 1.35106; G12= 0.83106; ν12= 0.3;tply = 0.005;

ν21=E2ν12

E1;

H∗ Reduced Stiffness matrix of angle plies∗L

Q11=E1

1− ν12ν21; Q22=

E21− ν12ν21

;Q12 = Q21 =E2ν12

1− ν12ν21; Q66= G12;

Q1−2 =88Q11, Q12, 0<, 8Q21, Q22, 0<, 80, 0, Q66<<;R = 0.765; RS = 0.265; L = 10; F= 500;

H∗Transformation about X axis∗Lmx= Cos@αD;tsigma= 981, 0, 0<, 90, HmxL2, 0=, 80, 0, mx<=;

tepsilon= 981, 0, 0<, 90, HmxL2, 0=, 80, 0, mx<=;

HQbarx−y,1 = Qbarx−y,4 = tsigma.Q1−2.tepsilonL êê MatrixForm;HQbarx−y,2 = Qbarx−y,5 = tsigma.Q1−2.tepsilonL êê MatrixForm;HQbarx−y,3 = Qbarx−y,6 = tsigma.Q1−2.tepsilonL êê MatrixForm;

H∗Transformation about Z axis∗Lm@θ_D:= Cos@θD;n@θ_D := Sin@θDTσ@θ_D := 99m@θD2, n@θD2, 2 m@θDn@θD=, 9n@θD2, m@θD2, −2 m@θDn@θD=,

9−m@θDn@θD, m@θDn@θD, m@θD2 − n@θD2==T∂@θ_D :=99m@θD2, n@θD2, m@θDn@θD=, 9n@θD2, m@θD2, −m@θDn@θD=,

9−2 m@θDn@θD, 2 m@θDn@θD, m@θD2 − n@θD2==JQx−y,1 = Qx−y,4 = Tσ@−θD.Qbarx−y,1.T∂@θD ê. θ →

π

4N êê MatrixForm;

JQx−y,2 = Qx−y,5 = Tσ@−θD.Qbarx−y,2.T∂@θD ê. θ →−π

4N êê MatrixForm;

HQx−y,3 = Qx−y,6 = Tσ@−θD.Qbarx−y,3.T∂@θD ê. θ → 0L êê MatrixForm;

tanphi =HR− RSL

L;

RX = R − HxtanphiL;T = 280;

H∗ Distance from Mid planes∗Lh= Table@i, 8i, −3, 3<D tply;

H∗ A B D matrices ∗Li

kjjjjA= ‚

k=1

6Qx−y,kHh@@k+ 1DD −h@@kDDL êêChop

y

{zzzz êê MatrixForm;

ijjjjB=1

k 2‚k=1

6Qx−y,kIh@@k+ 1DD2 − h@@kDD2M êê Chop

yzzzz êê MatrixForm;{

66

i

kjjjjDbs =

13

‚k=1

6Qx−y,k Ih@@k+ 1DD3 − h@@kDD3M êêChop

y

{zzzz êê MatrixForm;

Ad = A;Bd = B + RX Cos@αDA;Dd = Dbs + 2 RXCos@αD B + HRXL2 HCos@αDL2A;

A = RX ‡0

2 π

Ad αêê Chop;

i

k

jjjjjB = RX‡0

2 π

Bd αy

{

zzzzz êê Chopêê MatrixForm;

D = RX‡0

2 π

Dd αêê Chop;

Ia= Inverse@AD êê SimplifyM êê MatrixForm

δ = F ‡0

10

a@@1, 1DD x

H∗δ= F a@@1,1DDtanalpha

LogA RR−x tanalpha

Eê.x→10∗L

d= Inverse@DD;u= Apart@d@@3, 3DDD;φ = TChop Sum Integrate u i , x, 0, 10 , i, Length u@ @ @ @@ DD 8 <D 8 @ D<DD

67

APPENDIX E

ANSYS BATCH FILE FOR THE GENERATION OF THE FINITE ELEMENT MODEL

68

!* COMPOSITE TAPERED TUBE MADE OF 1M6/3501-6 WITH LAY-UP [45/-45/0]S /PREP7 !* MP,EX,1,22.8E6 MP,EY,1,1.35E6 MP,EZ,1,1.35E6 MP,PRXY,1,0.3 MP,PRYZ,1,0.3 MP,PRXZ,1,0.3 MP,GXY,1,0.83E6 MP,GYZ,1,0.83E6 MP,GXZ,1,0.83E6 !* ET,1,SOLID46 !* KEYOPT,1,2,0 KEYOPT,1,1,0 KEYOPT,1,3,0 KEYOPT,1,4,0 KEYOPT,1,5,0 KEYOPT,1,6,0 KEYOPT,1,8,1 KEYOPT,1,9,0 KEYOPT,1,10,0 !* !* LAYER 1 R,1 RMODIF,1,1,1,0,0,0 RMODIF,1,7,0 !* RMODIF,1,13,1,45,0.005 !* !* LAYER 2 R,2 RMODIF,2,1,1,0,0,0 RMODIF,2,7,0 !* RMODIF,2,13,1,-45,0.005 !* !* LAYER 3 R,3 RMODIF,3,1,1,0,0,0 RMODIF,3,7,0 !* RMODIF,3,13,1,0,0.005

69

!* !* LAYER 4 R,4 RMODIF,4,1,1,0,0,0 RMODIF,4,7,0 !* RMODIF,4,13,1,45,0.005 !* !* LAYER 5 R,5 RMODIF,5,1,1,0,0,0 RMODIF,5,7,0 !* RMODIF,5,13,1,-45,0.005 !* !* LAYER 6 R,6 RMODIF,6,1,1,0,0,0 RMODIF,6,7,0 !* RMODIF,6,13,1,0,0.005 !* !* ET,2,MASS21 !* !* KEYOPT,1,1,0 KEYOPT,1,2,0 KEYOPT,1,3,0 !* !* R,7,1,1,1, , , , !* !*CREATING KEYPOINTS !* *DO,I,1,7,1 K,I*10+1,0,0.75+(I-1)*0.005,0 K,I*10+2,10,0.25+(I-1)*0.005,0 *ENDDO !* !*CREATING AREAS USING KEYPOINTS !* *DO,I,1,6,1

70

A,11+(I-1)*10,12+(I-1)*10,12+(I*10),11+(I*10) *ENDDO AGLUE,ALL K,13,0,0,0 K,14,10,0,0 !* !* ROTATE AREAS TO FORM VOLUMES !* VROTAT,ALL,,,,,,13,14,,2 VGLUE,ALL !* !* CHANGING ATTRIBUTES OF VOLUMES TO LAYER PROPERTIES *DO,K,1,6,1 VSEL,S,VOLU,,K VATT,1,K,1,0 *ENDDO *DO,K,1,6,1 VSEL,S,VOLU,,K+6 VATT,1,K,1,0 *ENDDO !* !*DEFINING ELEMENT SIZES FOR LINES !* LSEL,S,LENGTH,,0.005,,,0 LESIZE,ALL,,,1 CM,THICKNESSLINE,LINE LSEL,S,LENGTH,,10.012,,,0 LESIZE,ALL,,,40 CM,LENGTHLINE,LINE ALLSEL,ALL,LINE LSEL,U,LENGTH,,0.005,,,0 LSEL,U,LENGTH,,10.012,,,0 LESIZE,ALL,,,12 !* !* CREATING NODES ON KEYPOINTS 13 AND 14 !* NKPT,,13

71

NKPT,,14 !* !* CREATING A MASS21 ELEMENT FOR MASTER NODE !* TYPE,2 REAL,7 E,2 !* !* MESHING THE VOLUMES !* VSEL,S,VOLU,,ALL VPLOT VMESH,ALL

72

REFERENCES

[1] Reddy, J N. (2003). Mechanics of Laminated composite plates and shells - Theory

and Analysis. CRC Press.

[2] Chan, W. S. and Demirhan, K. C., (1997). “A Simple Closed-Form Solution of

Bending Stiffness for Laminated Composite Tubes”, Journal of reinforced plastics and

composites, Volume 19, Issue 4, 2000, Pages 278-291

[3] Lin, C.Y. and Chan, W.S., (2003). “A simple analytical method for analyzing

laminated composite elliptical tubes”, Proceedings of 17th technical conference,

American Society of composites, October 2002.

[4] Ren, J. G. (1995). “Analysis of laminated circular cylindrical shells under

axisymmetric loading”, Composite Structures, Volume 30, Issue 3, 1995, Pages 271-280

[5] Tafreshi, A. and Bailey, C.G., (2007). “Instability of imperfect composite cylindrical

shells under combined loading”, Composite Structures, Volume 80, Issue 1, September

2007, Pages 49-64

73

[6] Vaziri, A., (2007). “On the buckling of cracked composite cylindrical shells under

axial compression”, Composite Structures, Volume 80, Issue 1, September 2007, Pages

152-158

[7] Yan, W., Ying, J. and Chen, W.Q.,(2007). “The behavior of angle-ply laminated

cylindrical shells with viscoelastic interfaces in cylindrical bending”, Composite

Structures, Volume 78, Issue 4, June 2007, Pages 551-559

[8] Gummadi, L. N. B. and Palazotto, A.N., (1998). “Progressive failure analysis of

composite cylindrical shells considering large rotations”, Composites Part B, Volume 29,

Issue 5, September 1998, Pages 547-563

[9] Correia, P. I. F., Soares, M. C.M. and Herskovits, J (2003). “Analysis of Laminated

Conical Shell Structures Using Higher Order Models”, Composite Structures, Volume

62, Issue 3, 2003, Pages. 383-390

[10] Liu, R.H (1996). “Non Linear Buckling of Symmetrically Laminated, Cylindrically

Orthotropic, Shallow Conical Shells considering Shear”, International Journal of Non-

Linear Mechanics, Volume 31, Issue 1, January 1996, Pages 89-99

[11] Goldfield, Y., (2007). “Elastic buckling and imperfection sensitivity of generally

stiffened conical shells”, AIAA Journal, Volume 45, Issue 3, March 2007

74

[12] Raju, B. B., Chandra, R. and Subba Rao, M., (1978). “Transient Temperatures in

Laminated Composite Conical shells due to Aerodynamic Heating”, AIAA Journal,

Volume 16, Number 6, 1978

[13] Ghosh, A. and Sinha, P. K., (2005). “Initiation and propogation of damage in

laminated composite shells due to low velocity impact”, International Journal of

Crashworthiness, Volume 10, Issue 4, Pages 379-388

[14] Mahdi, E., Hamouda, A.M.S., Sahari, B.B. and Khalid, Y.A., (2002). “Effect of

Material and Geometry on Crushing Behavior of Laminated Conical Composite Shells”,

Applied Composite Materials, Volume 9, Issue 5, September 2002, Pages 265-290

[15] Movsumov, E. A., and Shamiev, F. H. (2006). “Yield Condition for Circular

Cylindrical Shells Made of a Fiber Reinforced Composite”, Mechanics of Composite

Materials, Volume 42, Issue 5, September 2006, Pages 459-466

[16] Naidu, N.V.S. and Sinha, P.K. (2005). “Nonlinear finite element analysis of

laminated composite shells in hygrothermal environments”, Composite Structures,

Volume 69, Issue 4, August 2005, Pages 387-395

[17] Tafreshi, A., (2004). “Efficient modeling of delamination buckling in composite

cylindrical shells under axial compression”, Composite Structures, Volume 64, Issues 3-

4, June 2004, Pages 511-520

75

[18] Bhimaraddi, A. and Chandrashekhara, K., (1992). “Three-dimensional elasticity

solution for static response of simply supported orthotropic cylindrical shells”,

Composite Structures, Volume 20, Issue 4, 1992, Pages 227-235

[19] Saggar, P. (2007). Experimental Study of Laminated Composite Tubes under

Bending, Master’s Thesis, University of Texas at Arlington, May 2007

76

BIOGRAPHICAL INFORMATION

Chethana Shankara Rao received her Bachelor’s degree from Visveswaraiah

Technological University, India in 2003. She began her Master’s program in Mechanical

engineering at U.T. Arlington from August 2004.

Her research interests include Stress analysis, structural dynamics, fracture

mechanics and composite materials. She has carried out various projects on structural

analysis and intends to work in this field soon after the master’s degree. She also intends

to pursue a PhD degree in the future.

Chethana Shankara Rao received her master’s degree in mechanical engineering

from U.T Arlington in May 2007.

77